Data-Dependent Coresets for Compressing Neural Networks with Applications to Generalization Bounds.
Cenk BaykalLucas LiebenweinIgor GilitschenskiDan FeldmanDaniela RusPublished in: ICLR (Poster) (2019)
Keyphrases
- data dependent
- generalization bounds
- neural network
- uniform convergence
- pattern recognition
- risk bounds
- data compression
- neural network model
- rademacher complexity
- empirical risk minimization
- back propagation
- multiscale
- fault diagnosis
- learning theory
- energy functional
- kernel machines
- upper bound
- image processing
- feature selection
- machine learning